Learning algorithms with reinforcement on Python

This book will help the reader master the algorithms of training with reinforcement (OP) and learn to realize them when creating self -learning agents.
In the first part, various elements of the OP, the scope of its application, the tools necessary for work in the OP environment are considered. The second and third parts are devoted directly to the algorithms. Among other things, the author shows how to combine Q-teaching with neural networks to solve complex problems, describes the methods of the gradient of the strategy, TRPO and PPO, which can increase productivity and stability, as well as the determinated DDPG and TD3 algorithms. The reader learns how the imitative training technique works, get acquainted with the research algorithms based on the upper trusting border (UCB and UCB1) and the ESBAS meta-algorithm.
The publication is intended for those who are interested in research in the field of artificial intelligence, applies deep training in the work or want to master training with zero. A prerequisite is Python"s language ownership at the working level